Post on 20-Jul-2020
Variability and Predictability of the Ocean Thermohaline Circulation
Jochem MarotzkeMax Planck Institute for Meteorology (MPI-M)
Centre for Marine and Atmospheric Sciences (ZMAW)Bundesstrasse 53, 20146 Hamburg, Germany
With contributions by Johanna Baehr, Felix Landerer, Johann Jungclaus, Maria Paz Chidichimo (MPI-M), Stuart Cunningham, Joël
Hirschi, Torsten Kanzow, Harry Bryden (NOCS, Southampton)
Overview
1. Setting the Scene2. Decadal Climate Variability and
Predictability3. Observations of THC Change in the
North Atlantic
1. Setting the Scene
Observed Sea Level (⇔ Surface Circulation)
Rio & Hernandez (2004)
m
Observed Global Ocean Circulation
U.S. National Research Council (NRC, 2002)Abrupt Climate Change – Inevitable Surprises
Sea Surface Temperature, 4 – 9 November 2002
NomenclatureGulf Stream:
Narrow boundary current off North American coast (Florida)Pacific has counterpart (Kuro-shio)Gulf Stream cannot collapse, as long as winds blow, continents exist, and the Earth rotates
Meridional Overturning Circulation (MOC): Total northward/southward flow, over latitude and depthCounterpart to MMC in the atmosphereThermohaline Circulation (THC): Part of MOC driven by heat & water exchange with atmosphereMOC is observable quantity; THC an interpretationOften used synonymously, not rigorously correctHere: Use THC when confident of interpretation, MOCwhen rigour is required
Meridional Overturning Circulation (MOC)
Jayne & Marotzke 2001
2. Decadal Climate Variability and Predictability
Decadal Climate Variability and WCRPDecadal variability crucial for both main objectives of WCRP:
to determine the predictability of climate – start decadal climate predictions as an initial-value problem (WCRP Strategic Framework)
to determine the effect of human activities on climate – need to filter out natural decadal variability
Arguably: Ocean processes enhance decadal predictabilityLonger timescales: Large heat capacity (e.g., winter mixed layers)
Longer timescales: Slower dynamical processes
Arguably: THC, rather than wind-driven circulation, enhances decadal climate predictability
THC more likely to be governed by slower oceanic processes
THC important for climatic influence and for predictability
Mechanisms of Decadal THC Variability
Modelling THC variability far more mature than observations – worrisome!Still not clear whether coupled mode (Timmermann et al. 1998) or stochastically driven (Delworth et al. 1993), possibly enhanced by damped (Griffies and Tziperman 1995) or self-sustained (Marotzke 1990, Weaver and Sarachik 1991) ocean modesMainly heat flux-driven as a robust result?Effect of decadal THC variations on European climate seen in models (Pohlmann and Keenlyside 2004, Sutton and Hodson 2005) and observations (Czaja and Frankignoul 2002)
Simulated Atlantic MOC
Roeckner et al. (2006)
Ice Age or Hothouse – Which Is It to Be?
PM September 2004 Title
Can We Predict a Possible THC Downturn?
Are all important processes included in the models?Influence of Greenland meltwater on THC stability (not included in the protocol for IPCC AR4 runs)
Necessary for prediction: continuous observation of the very quantity that is to be predicted
Starting point of the proposal to UK NERC to establish the RAPID programme (Marotzke et al., 2000)
“Greenland Melts,” MOC Strength
A1B; A1B+0.03 Sv; A1B + 0.09 Sv
Jungclaus et al. (2006)
20CC + 0.09 Sv
20CC
3. Observations of MOC Change in the North Atlantic
North Atlantic Circulation
Quadfasel (2005)
Observations of Change Related to the MOC
Dickson et al. (2002), Curry et al. (2003): Freshening in northern North Atlantic over last 4 decades (hydrography)Hansen et al. (2001): Reduction in overflows (hydrography + hydraulic control theory)Häkkinen and Rhines (2004): Slowdown of subpolar gyre surface circulation, 1992-2003 (altimetry)All high-profile papers (Nature, Science); public discussion seemed to imply a corresponding weakening of MOCBUT: No indication these measures are valid proxies of MOC – on the contrary (HadCM3; ECHAM5/MPI-OM):
Wu et al. (2004): Freshening coincides with stronger MOCLanderer et al. (2006): No correlation subpolar gyre strength-MOC
Baroclinic gyre transport
Δ_Sea levelBermuda -Labrador Sea
MOCAtlantic 30°N
Correlation
Control (Grey) & IPCC 20C + A1B Simulations (Blue)
Landerer et al. (2006)
2000
Nature, 1. December 2005
Bryden et al. (2005): Weakening of MOC at 25°N by 30%, 2004 relative to 1957 (and relative to 1992)But: No changes in boundary currents, whether in subtropical (Baringer and Larsen 2001) or subpolar gyre (Schott et al. 2006)But: Why was the 1°C cooling expected with such an MOC slowdown (R. Wood, in Kerr 2005) not observed? But: Do 5 “snapshots” (Oct 1957, Jul/Aug 1981, Jul/Aug 1992, Feb 1998, April 2004) allow us to distinguish between trend and variability?
Simulated Atlantic MOC at 26°N
(σ = 1.7Sv)
Observations (duration 1 month each)
Johanna Baehr (2006)
(σ = 1.0Sv)
Observed vs. modelled variability
Baehr et al. (2006)
Detecting Modelled MOC Change
Baehr et al. (2006)
Observed vs. modelled variability
Baehr et al. (2006)
Feb. 2004: Continuous Observations Started
Schiermeier (2004)
Near Atlantic heat transport maximum – captures total heat transport convergence into North Atlantic
South of area of intense heat loss from ocean to atmosphere over Gulf Stream extension
MOC dominates heat transport (Hall & Bryden ‘82)
Heat transport variability dominated by velocity fluctuations (Jayne & Marotzke, 2001)
Florida Strait transport monitored for >20 years (now: Johns, Baringer, Meinen & Beal, Miami, collaborators)
5 modern hydrographic sections (‘57, ‘81, ‘92, ’98, ´04)
Why 26.5°N?
Light blue: Velocity determination fromdensitymeasurements
Green: Velocity determination fromwind measurements
Red: Florida Straittransportmeasurements withtelephone cable
Monitoring the Atlantic MOC at 26.5°N(Marotzke, Cunningham, Bryden, Kanzow, Hirschi, Johns, Baringer, Meinen, Beal)
Yellow: Uniform correction for massconservation
Hirschi (2005)
Monitoring the Atlantic MOC at 26.5°N(Marotzke, Cunningham, Bryden, Kanzow, Hirschi, Johns, Baringer, Meinen, Beal)
2004 mooring deployment:
Data recovery: April, May, Oct. 2005; March, May 2006
Monitoring the Atlantic MOC at 26°N(Marotzke, Cunningham, Bryden, Johns, Baringer, Beal)
Monitoring the Atlantic MOC at 26.5°N(Marotzke, Cunningham, Bryden, Kanzow, Hirschi, Johns, Baringer, Meinen, Beal)
April 2005
Oct. 2005
Waterfall Plot, Salinity vs. Temperature from Moored Profiler
Maria Paz Chidichimo (2006)
April 2005
Oct. 2005
Waterfall Plot of Potential Density from Moored Profiler
Maria Paz Chidichimo (2006)
Contributions to Integrated Transport Variability
Kanzow et al. (2006)
Mid-Ocean Geostrophic Transport Variability
Bryden et al. (2006)
(σ = 2.6Sv)
(σ = 1.8Sv)
(σ = 2.0Sv)
Time (Days) April 2004 to May 2005
ConclusionsGreenland meltwater only moderately destabilising for THC during the next two centuries
No valid proxy for MOC has yet been identified
Continuous observing system of Atlantic MOC has been put in place at 26.5°N.Observations show surprisingly strong high-frequency variability of the MOC“Observations” of MOC slowdown likely to be artefact of temporal subsampling of noisy system
Outlook
MOC time series needs to be continuedAlternative observing systems? Cheaper technologies (obviate moorings? Full-depth gliders?)Transfer to operational agencies after (likely) RAPID-WATCH phase ends in 2014Complementary locations (northern North Atlantic? South Atlantic?)
Development of MOC proxiesSimple proxies (e.g., SST, Latif et al. 2004)Multiproxies (ultimate multiproxy: ocean re-analysis)
Outlook
Decadal predictability of MOC and climateMove decadal predictability studies from pure modelling exercises into initialisation of global coupled models with observations, including global data assimilation (ocean & coupled re-analysis)Measurements of MOC, MOC proxies, quantities influenced by MOC crucial
Mechanisms of interdecadal MOC variabilityPicture still very unclear, but many groups work on itToo much focussed on pure modelling studies?Learn from ENSO theory to consider superposition of effects?
Thank you for your attention!